Upload
duongtram
View
217
Download
0
Embed Size (px)
Citation preview
The Gut Microbiome: Markers of Human Health, Drug Efficacy and Xenobiotic Toxicity
Biomarkers of Toxicity and Disease
Identification of microbiome-based biomarkers and challenges associated with
their application: Case studies from obesity, IBD, and cancer
25 June 2018 | Alexandria, VA
Emily Hollister
VP, Information Technology & Analytics
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20181
ORIGIN
Background – Diversigen is a commercial microbiome service provider powered by CMMR
HEADQUARTERS
PEOPLE
Houston, Texas - located at
the heart of the Texas
Medical Center
12 highly talented and
experienced individuals
with a total of 8 PhD’s
YEAR FOUNDED
2013
Operational since
December, 2014
MICROBIOME SERVICES
Metagenomic sequencing
pipelines | Germ-free
Project consulting
Bioinformatics
Microbiome Exploration | Microbial
ecology modeling and dissection
Therapeutic development | Policy and
outreach | Education | Translation
Centre for Metagenomics
and Microbiome Research
(CMMR)
DIVERSIGEN
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20182
Patient Stratification
Metagenomics plays a key role in therapeutic and diagnostic development
Probotic / antibiotic
Diagnostic
.
Compare microbiota in healthy and ill
subjects
ID organisms / genes associated with health /
disease
Develop applications /
solutions
▪ Clinical trials
▪ Personalized medicine
Community
(100’s of strains,
undefined composition)
Consortium
(defined composition of more
than one strain, which together
perform a function of interest
Single strain
(one strain, pure isolate)
Bioactive
(molecule produced by strain
that mediates effect of host)
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20183
Microbiome-based biomarkers
•What are they? Do they exist?
•What are the challenges associated with their application?
•Defining healthy has turned out to be difficult
•Consensus/lack of across multiple studies – sources of bias, evolution
of microbiome science/understanding
•They don’t have to be bacterial – although many purported biomarkers
are
•The majority of the data are 16S-based
•16S lacks the resolution that we’d prefer to have (i.e., better than
genus)
•The majority of data is stool-based
•good for broad applicability and user acceptance; may not reflect
biology at the disease interface. But…..at the end of the day, does it
matter?
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20184
Obesity and the gut microbiome
• Summarizing the literature:
• Data from animal studies is very compelling
• Data from human studies is somewhat murky
• Consensus across studies is limited
• Multiple potential pitfalls
Ley et al 2006; Turnbaugh et al 2006
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20185
CARB-R
T0T0T2T3T1T0T1T2T1T2T0T2T3T1T2T3T1T1T3T3T0T0
FAT-R
T0T0T0T2T1T0T1T3T2T3T1T2T1T3T0T3T0T2T1T3T0T1T2T3
a
Change in body weight (%)
Change in B
acte
roid
ete
s
abundance (
%)
-50
R2 = 0.8
R2 = 0.5
-25-20-15-10
0
5
10
15
20
25
30 CARB-RFAT-R
c
Perc
enta
ge o
f to
tal
sequences
Weeks on diet0 12 26 52 Lean
20
40
60
80
100
Firmicutes Bacteroidetesb
Ischnura hastata
Varanus
MICROBIAL ECOLOGY
Human gut microbes associated with obesity
ob ob
ob ob
+/+
Figure 1 | Correlation between body-weight loss and gut microbial ecology. a, Clustering of 16S ribosomal RNA gene sequence libraries of faecal microbiota for each person (in different colours) and time point in diet therapy (T0, baseline; T1, 12 weeks; T2, 26 weeks; T3, 52 weeks) in the two diet-treatment groups (fat restricted, FAT-R; carbohydrate restricted, CARB-R), based on UniFrac analysis of the 18,348-sequence phylogenetic tree. b, Relative abundance of Bacteroidetes and Firmicutes. For each time point, values from all available samples were averaged (n was 11 or 12 per time point). Lean-subject controls include four stool samples from two people taken 1 year apart, plus three other stool samples6. Mean values ± s.e. are plotted. c, Change in relative abundance of Bacteroidetes in subjects with weight loss above a threshold of 2% weight loss for the CARB-R diet and 6% for the FAT-R diet.
Phillip C. Watts*, Kevin R. Buley†,
Stephanie Sanderson†, Wayne Boardman‡,
Claudio Ciofi§, Richard Gibson‡
*School of Biological Sciences, University of
Liverpool, Liverpool L69 7ZB, UK
e-mail: [email protected]
†North of England Zoological Society, Chester
Zoo, Upton-by-Chester CH2 1LH, UK
‡Zoological Society of London, Regents Park,
London NW1 4RY, UK
§Department of Animal Biology and Genetics,
University of Florence, 50125 Florence, Italy
1. Ciofi, C. & De Boer, M. Herpetol. J. 14, 99–107 (2004).
2. Ciofi, C. & Bruford, M. W. Mol. Ecol. 7, 134–136 (1998). 3. White, M. J. D. Animal Cytology and Evolution (Cambridge
University Press, Cambridge, UK, 1973). 4. Avise, J. C., Quattro, J. M. & Vrijenhoek, R. C. in
Evolutionary Biology (eds Hecht, M. K., Wallace, B. & MacIntyre, R. J.) 225–246 (Plenum, New York, 1992).
5. Groot, T. V. M., Bruins, E. & Breeuwer, J. A. Heredity 90, 130–135 (2003).
6. Lenk, P., Eidenmueller, B., Staudter, H., Wicker, R. & Wink, M. Amphibia-Reptilia 26, 507–514 (2005).
7. Saccheri, I. J. et al. Nature 392, 491–494 (1998). 8. Spielman, D., Brook, B. W. & Frankham, R. Proc. Natl Acad.
Sci. USA 101, 15261–15264 (2004). 9. Sherratt, T. N. & Beatty, C. D. Nature 435, 1039–1040
(2005). 10. Halverson, J. & Spelman, L. H. in Biology and Conservation
of Komodo Dragons (eds Murphy, J. B., Ciofi, C., de La Panouse, C. & Walsh, T.) 165–177 (Smithsonian Institution, Washington DC, 2002).
Supplementary information accompanies this communication on Nature’s website. Received 4 October; accepted 16 November 2006.Competing financial interests: declared none. doi:10.1038/nature4441021a
1022
NATURE|Vol 444|21/ 28 December2006BRIEF COMMUNICATIONS
Nature Publishing Group ©2006
Conferring obesity through FMT
Alang and Kelly 2015; Cox et al 2014
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20186
Microbiome alterations in are subtle and only partially reproducible
Lopez-Contreras et al 2017 Pediatric Obesity; Hollister et al 2018 Childhood Obesity; Riva et al 2017 Environmental Microbiology
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20187
Microbiome alterations in obesity are inconsistent
Sze and Schloss 2016
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20188
IBD: The interaction of genetics, the immune system, and environment
• IBD includes Crohn’s Disease and Ulcerative Colitis
• IBD affects 1.4M Americans, with increasing incidence
in recent decades
• Greater incidence in Western nations and countries
experiencing demographic/economic development
• Observed at greater rates in children immigrating to
high(er) prevalence countries
• Concordance <50% in monozygotic twins
• Direct cost of treatment in the US represents an
economic burden of ~$4 billion annually (UC)
• Greater risk for developing colorectal cancer
Danese, S. & Fiocci, C. (2011). NEJM; Kornbluth A, Sachar DB (2004). Am. J. Gastroenterol;Cohen et al. (2010). Aliment Pharmacol Ther; Etchvers et al. 2008. World J Gastroenterol; Talley et al. (2011). Am J. Gastroenterol
No data Low-medium incidence High incidence
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 20189
IBD: Taxonomic patterns and trends
Taxonomic patterns:
Imbalances/dysbiosis commonly described in the
literature – largely 16S based:
• CD: ⇩ Roseburia, ⇩ Faecalibacterium
prausntizii, ⇧ Ruminococcus gnavus
• UC: not as clearly defined
• ⇧ sulfate-reducing Proteobacteria have
been reported
• ⇩ Roseburia, ⇩ Ruminococcaceae,
Gevers et al 2014 Cell Host Microbe (CD)Zhou et al 2018 mSystems
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201810
IBD: Functional metagenomic patterns and and trends
Morgan et al (2012) Genome ResearchGevers et al 2014 Cell Host Microbe
Functional patterns:
Fewer studies have used functional metagenomics.
• Gevers et al 2014: Species increased in CD
contributed enrichment of glycerophospholipid
and lipopolysaccharide metabolism
• Morgan et al 2012: Increased metabolism of
sulfur-containing amino acids in IBD; shifted
metabolism toward the degradation of mucin
• Alterations suggest an environment
characterized by inflammation and oxidative
stress
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201811
Does consensus exist with respect to IBD biomarkers?
Zhou et al 2018 mSystems
Consensus?
To some degree, yes:
• Studies have repeatedly
identified many of the same taxa
(16S, WGS)
• ⇧ Enterobacteriaceae
• ⇧ Pasteurellaceae
• ⇧ Veillonellaceae
• ⇧ Fusobacteriaceae
• ⇩ Faecalibacterium
• ⇩ Roseburia
• ⇩ Clostridiales
• Unclear roles: Causal,
contributory, coincident,
consequential
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201812
IBD: How might microbiome-based biomarkers be used?
Zhou et al 2018 mSystems
How might biomarker be used?
1. Early diagnosis
2. Monitoring of disease status/severity
3. Evaluating treatment success
4. Monitoring progression otherwise (e.g., toward relative health or other disease states)
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201813
CRC & IBD: Risk association
• Increased risk of CRC in IBD recognized in late 80s
• Reported risk factors for CRC include
• extensive disease
• young age at diagnosis
• family history of CRC
• co-existing primary sclerosing cholangitis
• persistent inflammation of the colon
• CRC in IBD carries a 2-fold increase in mortality compared with sporadic cancers.
• No major differences between UC and CD except for a more advanced tumor
stage at the time of diagnosis in CD patients.
• With IBD, the presence of CRC is often not known prior to surgery
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201814
Inflammation in CRC
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201815
Gao et al 2017 American Journal of Pathology
Overall survivorship influenced by gene expression
Histidine decarboxylase Histamine H2 receptor
Inflammation in CRC
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201816
Gao et al 2017 American Journal of Pathology
Model without chemical trigger
Model with chemical trigger
Model with chemical trigger and microbe
Model with chemical trigger and ko mutant microbe
Microbiome attenuation by addition of one organism
Colorectal cancer and the gut microbiome
• First CRC microbiome associations: adherent-invasive Escherichia coli is
associated both with IBD and CRC
• Modern metagenomic techniques have yielded conflicting results
• Most studies find associations
• Different taxa are associated with CRC presence/absence, severity,
disease progression, and survivorship
• Will microbiome be able to answer the question alone?
•What about in high risk populations?
• CRC and the gut microbiome — Where are we with respect to describing it
and identifying actionable biomarkers?
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201817
Rubin et al (2012) Front Immunol
Disparate CRC associated microbiota
Weir et al (2013) PLoS One; Zeller et al (2014) Molecular Systems Biology; Wang et al (2014) ISME; Hannigan (2017) bioRxiv
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201818
When processed similarly concordance in associations
Shah et al 2017 Gut
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201819
…mostly.
Improvement in sensitivity by microbiome + risk factors
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201820
Baxter et al (2016) Genome Medicine; Zellar et al (2014) Molecular Systems Biology; Zackular et al (2014) Cancer Prev Res
gFOBT, BMI, Wif1 methylation, fecal immunochemical test, viral clades…
Addition of metagenomic data
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201821
Hannigan (2017) bioRxiv
• 16S
• WGSBacterialViral
• Depth of sequencing
Multiple studies processed separately
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201822
Shah et al 2017 Gut
Multiple studies processed together
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201823
Shah et al 2017 Gut
$$
Microbiome-based biomarkers
Where do we go from here?Ideal biomarkers should be able to:
• Identify disease at/prior to onset; differentiatebetween stages of disease (i.e., disease progression and/or response to treatment)
• Be accessible and affordable to facilitate broad acceptance for large scale screening
• Work in high-risk populations (e.g., T2DM, IBD are risk factors for CRC)
• Distinguish between disease of interest and other potentially microbiome-mediated conditions
• Be robust to lifestyle factors such as OTC or prescription drug consumption
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201824
*XX
Performing metagenomics at scale has many advantages
Studying multiple disease cohorts can facilitate dissection of disease mechanisms and lead to better
diagnostics and therapeutics
Universal factors:
• Common exposures
• Common genetic risks
Unique regional factors:
• Diet
• Race/ethnicity
• Access to health care
• Environment
* Universal target discovery * Personalized target
discovery
Cohort 1
Cohort 3
Cohort 5Cohort 4
Cohort 2
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201825
Future outlook – efforts to accelerate and improve microbiome drug development
Utilize libraries of microbiome
samples and dataAdopt industry standards e.g.
CLIA, GLP
Improve study design and share
clinical knowledge & insights
Grow functional data (WGS,
imputed analysis, IgASeq, RNAseq,
etc)
Increase Black Box Methods to
identify/classify microbiomesFacilitate the shift from
association to causality
Access to sequencing resources
and comprehensive platforms
Incorporate machine learning for
predictive and association analysis
Diversigen Inc. | HESI Microbiome Workshop| 25-26 June, 201826